Rendering large-scale terrain models and positioning objects in relation to 3D terrain

Abstract

Computer generated 3D graphics have been commonplace in computing since the early 1990's. However, most 3D scenes have focused on relatively small areas such as rooms or buildings. Rendering large scale landscapes based on 3D geometry generally did not occur because the scenes generated tended to use up too much system memory and overburden 3D graphics cards with too many polygons. However, there are applications where the terrain is critical and needs to be rendered properly such as cartography and military simulation. This thesis is focused on methods of rendering terrain for such applications. The data used to build terrain geometry typically comes from elevation postings taken from surveys of the terrain. This thesis does not focus on collecting this data nor does it compare various sources of terrain data. Instead, this thesis is about taking elevation data, producing a rendered 3D scene, and placing objects within the scene relative to the terrain. Having these capabilities makes many military and cartographic applications possible. Some military applications include displaying the results of computer simulations in 3D, planning operations using a 3D landscape, and rehearsing operations in 3D. The military does have some tools that can be used today for these actions, but the tools are typically proprietary and expensive. This thesis is focused on using and extending open source tools for 3D terrain rendering. The result is tools that can be freely used, studied, and expanded by anyone without licensing costs.